The long term goal of this research project is to understand the molecular mechanism of force production through 3-D visualization of crossbridge states in situ in muscle. The research focuses on the structure of the large waterbug Lethocerus sp. because its filament lattice is the best ordered of all known muscle types thereby making it an excellent candidate for 3-D imaging as well as facilitating the trapping of many crossbridges into similar structures. Specimen preparation emphasizes rapid freezing and freeze substitution which traps molecular motions with millisecond time resolution. Crossbridges during isometric and stretch activated contraction are emphasized and complemented by states stabilized with nucleotide analogues such as ADP-AIF4 and ADP-Vanadate +/- Ca2+. Ramp stretched and slack rigor muscle will be compared to determine structural changes possible for strongly bound 2-headed crossbridges. Response of crossbridges to mechanical perturbations such as quick stretch and release and ramp stretch will be trapped by fast freezing and imaged in 3-D. Structures observed by 3-D electron microscopy will be correlated with mechanical measurements made prior to freezing. Reconstruction work will focus on 30-50 nm sections which yield the highest detail on crossbridge structure. We will utilize electron tomography to obtain 3-D images of muscle crossbridges without spatial averaging and use 3-D correspondence analysis to identify groups of similarly structured crossbridges for subsequent averaging to improve the signal to noise ratio. Continued refinements of our unique tomographic method are proposed in order to increase resolution, improve the reconstructions and the rate at which 3-D maps can be produced. Atomic models based on the crystal structures of actin and myosin will be built and quantitatively fit within the envelope of the reconstruction. These atomic models will be used to make predictions of domain orientations and distances and compared whenever possible with corresponding data from spectroscopy.